Liquid sample supplied to an analysis device in the form of droplets is usually provided using a nebulizer to generate an aerosol. Analysis devices which utilize such droplets include ionization and/or excitation sources such as microwave induced plasma (MIP), inductively coupled plasma (ICP) and flames. The analysis devices provide spectrometers which perform MIP and ICP optical emission spectrometry (OES), MIP and ICP mass spectrometry (MS), atomic absorption spectrometry (AA) and atomic fluorescence spectroscopy (AFS). Typically the sample-containing liquid is formed into a stream of droplets using a nebulizer utilizing a stream of inert gas such as argon. Nebulizers produce droplets with a wide range of sizes. However where the analysis device utilizes a plasma or flame to dissociate and excite or ionize the sample, as both plasmas and flames are inefficient at dissociating large droplets, a spray chamber is usually placed between the nebulizer and the torch so as to exclude large droplets from the sample stream entering the analysis device. The spray chamber filters the stream of droplets by causing the flow to follow a tortuous path such that the larger droplets impinge upon surfaces in the spray chamber and are drained away, smaller droplets being carried by the flow of gas into the torch. In the cases of ICP-OES and ICP-MS it is well known that only 1-2% of the nebulized sample-containing liquid is in the form of sufficiently small droplets suitable for processing within the torch, and that this form of sample introduction is therefore inefficient.
Alternative methods of producing a stream of sample droplets include the use of continuous fluid jet micro-droplet generator (G. M. Hieftje and H. V. Malmastadt, Analytical Chemistry, Vol. 40, pp. 1860-1867, 1968) and vibrating orifice monodisperse aerosol generator (H. Kawaguchi et al., Spectrochimica Acta, vol. 41 B, pp. 1277-1286, 1986, T. Nomizu et al., Journal of Analytical Atomic Spectrometry, vol. 17, pp. 592-595, 2002), The ability to produce droplets one at a time and thereby more completely control the droplet ejection process—so-called “droplet-on-demand” techniques—have long been seen as desirable. Where the droplet generator is of a type in which the droplet generation apparatus enables a single droplet to be ejected in response to a control signal, the droplet generator is one of a class of generators termed droplet-on-demand generators. An early generator with this capability designed principally for inkjet printing was a piezoelectrical droplet generator (U.S. Pat. No. 3,683,212). Such a droplet generator was employed to create a stream of droplets containing sample material, the droplets being passed through an oven so as to make the droplet evaporate to complete or partial dryness before injection into an ICP in order to study oxide ion formation (J. B. French, B. Etkin, R. Jong, Analytical Chemistry, Vol. 66, pp. 685-691, 1994). This coupling of the piezoelectric droplet generator and oven was termed the monodisperse dried microparticulate injector (MDMI) and such systems have been used in other studies (J. W. Olesik and S. E. Hobbs, Analytical Chemistry, vol. 66, pp. 3371-3378, 1994; A. C. Lazar and P. B. Farnsworth, Applied Spectroscopy, vol. 53, pp. 457-470, 1999; A. C. Lazar and P. B. Farnsworth, Applied Spectroscopy, vol. 51, pp. 617-624, 1997). Use of the piezoelectric droplet generator without the desolvation in an oven has been successfully implemented as a sub-nanoliter sample introduction technique for Laser-Induced Breakdown Spectroscopy and Inductively Coupled Plasma Spectrometry (S. Groh et al., Analytical Chemistry, vol. 82, pp. 2568-2573, 2010; A. Murtazin et al., Spectrochimica Acta, vol. 67B, pp. 3-16, 2012).
All these droplet generation devices require liquid sample to be fed into an enclosed volume within the droplet generation device. Typically sample is prepared and stored in vessels, and the vessels are usually stored in an array close to the analysis device, so that the vessels may be accessed by an autosampler. The autosampler positions a take-up tube within one of the vessels, and sample is drawn into the tube and transported into the droplet generator using suction. Hence the sample-containing liquid comes in contact with the take-up tubing and with the internal surfaces of the droplet generator. Once the sample take-up is complete, the autosampler withdraws the take-up tubing from the sample-containing vessel and moves it to a vessel containing wash solution. Wash solution is drawn into the take-up tubing and into the droplet generator and flushed to waste in order to wash out the remains of the previous sample before the next sample is admitted. For all the droplet generation devices described above, whether or not an autosampler is utilized, means such as tubing to transfer sample-containing liquid from a storage vessel is required and the droplet generator itself presents exposed surfaces to the sample-containing liquid.
Due to the increasingly routine use of spectrometry, sample throughput has become one of the most important requirements as often it is this which ultimately determines the cost-per-analysis in routine applications. With the increased sensitivity of instrumentation and automated sample handling, sample throughput is largely limited not by the sample introduction or analysis time but rather by memory effects caused by deposition of material from the previous sample on components of the sample introduction system and spectrometer. Due to the increased sensitivity of the spectrometers and their ultimate detection limits, material deposited upon the sample introduction system is gradually washed away during the “wash” cycle described above, and typically at least 40-60 seconds is needed after each sample to reduce memory effects below an acceptable threshold. In addition, the time to transport liquid from a containment vessel to the droplet generator may be significant, adding time both for sample uptake and wash solution uptake.
Development of instrumentation has increased the sensitivity of analysis devices and frequently sample solutions require dilution. Various methods for automatic dilution of samples have been devised (as described for example in U.S. Pat. No. 7,998,434). In order to monitor and correct for variations in accuracy, internal standards are often used. Both dilution and addition of standards requires the mixing of liquids prior to introduction to the analysis device. With all the droplet generation devices above, typically the liquids to be mixed are either mixed within a vessel prior to take-up, or are mixed at a location between the vessels containing the liquids and the droplet generator. As such, additional liquid handling devices or process steps are required, and additional vessels or separate mixing devices are required. Any mixing devices and associated liquid containment conduits must also be washed out prior to their next use.
Acoustic droplet ejection systems have been developed utilising a phenomenon first reported by R. W. Wood and A. Loomis in 1927 [Philiosophical Magazine, 4 (22), 417-436]. Acoustic energy emitted from a transducer can be converted to kinetic energy in a liquid. If acoustic energy is focused near a free surface of the liquid, droplets may be ejected from the surface of the liquid, the droplet size scaling inversely with the frequency of the acoustic energy. Droplet volumes from ˜20 μl to 2 μl and droplet ejection rates of hundreds of droplets per second may be produced. Unlike other droplet ejection devices, no contact between the sample liquid and the droplet ejector or sampling apparatus such as nozzles, pipette tips or pin tools occurs. Prior art acoustic droplet ejectors have been used to eject droplets upwards from well plates to be deposited onto solid surfaces or receiving plates located immediately above the well plates. Hence droplets are transferred from containment vessels onto receiving vessels in relatively close proximity.